On 5/13/20 11:40 PM, Michal Hocko wrote:
On Wed 13-05-20 23:11:40, Naresh Kamboju wrote:
While running libhugetlbfs fallocate_stress.sh on stable-rc 5.4 branch kernel on arm64 hikey device. The following kernel Internal error: Oops: crash dump noticed.
Is the same problem reproducible on vanilla 5.4 without any stable patches?
Or, an earlier version of 5.4-stable? Nothing in the changelog for 5.4.41 looks related to this issue. There was an arm specific hugetlb change "arm64: hugetlb: avoid potential NULL dereference", but that is pretty straight forward.
I'm guessing this may not reproduce easily. To help reproduce, you could change the #define FALLOCATE_ITERATIONS 100000 in .../libhugetlbfs/tests/fallocate_stress.c to a larger number to force the stress test to run longer.